Virtual Agent Standard Flows for Help Center

[Jonne Riikonen]

Hello Community

We are exploring the ways to integrate the long-awaited AI based Virtual Agent to our Help Center.

While there are plenty of matters to still attend, we are making good progress but came across somewhat of a weird detail.

As base info, we've got two Projects in our Help Center, one contains all the customer requests (separated by Organizations) and the other one is a customer specific Knowledge Base with access limited to one Organization only (not ideal but the only way to do a limited-access KB for the time being, as far as we know).

Now, to the matter at hand.

We've successfully enabled the Virtual Agent and it also manages to bring our some relevant answers with AI. However, we believe that it would make sense to insert a disclaimer for the customers which states that if they take any actions based on the AI answers, we're not actually liable for any damages it may cause (especially if the AI answer misleads the customer to wrong way... we cannot affect that).

To do this, the simple way seemed to alter the Standard flow / Greet by simply adding a few notes about the terms & conditions and stating that by continuing they accept them.

We did this in both of our Service Projects, and it worked well, unfortunately though, if we access the Virtual Agent chat from the Help Center front page, it does not use those same flows but something else. We've looked everywhere for those but cannot find them. We are now assuming those are hard-coded beyond our access, which actually renders the whole thing a little silly since we would not be sure if it would know to match the taught Intents either since it seems to be working in a different layer.

Question:

Is there a place to access those "global" or Help Center level Standard flows?

If the answer is "No", question2: Is there a place to disable the Virtual Agent from the Help Center but keeping it on the separate Portals? EDIT: This seems to be possible, but renders the user experience less ideal.

We hope that a solution exists, because unfortunately it seems that without addressing this issue we won't be able to properly deploy the Virtual Agent to our Help Center.

[Christopher Yen]

Hi @Jonne Riikonen 

Unfortunately it seems like there are no intents on the Help Center level virtual agent but it's supposed to try to match intents from project levels according to this documentation 

If projects linked to your help center already use the virtual service agent and have any intents set to Live, the virtual service agent in the help center will try to match these intents when responding to customer queries. You can’t create intents for a help center – these must be created at the project level.

Deactivate the virtual service agent in your help center

You need to be a site admin or organization admin to activate or deactivate the virtual service agent.

To deactivate the virtual service agent in your help center:

1. From your service project, select Channels, and then Portal.

2. Select your avatar in the top right hand corner.

3. Select Virtual service agent.

4. Select Settings.

5. Use the toggle next to Virtual service agent to deactivate the virtual service agent in your help center.

[Jonne Riikonen]

Thanks Christopher!

I actually found the way to disable the Virtual service agent from the Help Center front page soon after I wrote here (a classic... heh). Unfortunately this brings the user experience down a lot because one then has to choose one of the Portals to access the Agent.

This particular scenario for us is very inconvenient as we have targeted to have all our actual support tickets in the "main portal" and all the parallel portals are only for Knowledge Base use (either per product or per customer).

Having the possibility of triggering an "Interactive AI session" on the front page directly would be the ideal way forward since the search itself already provides potentially good matches to the customer query, given the customer has access to those articles (via specific portal access).

Maybe there's a way to work via Intents but it doesn't really resolve the issue we have.

I'll be looking deeper for a workaround.